Mujat Mircea, Chan Raymond, Cense Barry, Park B, Joo Chulmin, Akkin Taner, Chen Teresa, de Boer Johannes
Opt Express. 2005 Nov 14;13(23):9480-91. doi: 10.1364/opex.13.009480.
We introduce a method to determine the retinal nerve fiber layer (RNFL) thickness in OCT images based on anisotropic noise suppression and deformable splines. Spectral-Domain Optical Coherence Tomography (SDOCT) data was acquired at 29 kHz A-line rate with a depth resolution of 2.6 mum and a depth range of 1.6 mm. Areas of 9.6x6.4 mm2 and 6.4x6.4 mm2 were acquired in approximately 6 seconds. The deformable spline algorithm determined the vitreous-RNFL and RNFL-ganglion cell/inner plexiform layer boundary, respectively, based on changes in the reflectivity, resulting in a quantitative estimation of the RNFL thickness. The thickness map was combined with an integrated reflectance map of the retina and a typical OCT movie to facilitate clinical interpretation of the OCT data. Large area maps of RNFL thickness will permit better longitudinal evaluation of RNFL thinning in glaucoma.
我们介绍一种基于各向异性噪声抑制和可变形样条来确定光学相干断层扫描（OCT）图像中视网膜神经纤维层（RNFL）厚度的方法。光谱域光学相干断层扫描（SDOCT）数据以29 kHz的A线速率采集，深度分辨率为2.6μm，深度范围为1.6 mm。在大约6秒内采集9.6×6.4 mm2和6.4×6.4 mm2的区域。可变形样条算法分别根据反射率变化确定玻璃体-RNFL和RNFL-神经节细胞/内丛状层边界，从而对RNFL厚度进行定量估计。厚度图与视网膜的综合反射率图以及典型的OCT动态图像相结合，以促进对OCT数据的临床解读。RNFL厚度的大面积图将有助于更好地对青光眼患者的RNFL变薄情况进行纵向评估。
